I am currently designing a circuit where:

  • An LED driven by a constant current driver shines directly into a photodiode. (I have also tried without the op amp constant current circuit.)
  • The photodiode is operating in photovoltaic mode, into an op-amp.
  • The output of the op-amp is directly plugged into my headphones (I should AC couple but not in the schematics).


Now, what is strange is :

  • When the photodiode is covered, I can hear absolutely no noise (even when the LED is lit).
  • When I let my desk lamp shine into the photodiode, I can clearly hear the 60 Hz hum. Using a membrane partially covering the photodiode, I can hear sounds made by the vibration of the membrane.
  • When I approach the LED, I hear a very strong background high frequency white noise (some sort of hiss).

The LED is powered through DC (I would say reasonably clean DC), so where is this "optical noise" coming from? Is this the 1/f noise some papers talk about? Is there a driver circuit to avoid it? Alternative light sources?

EDIT: 21/11/23

First, thanks you all, it is very formative. It seems I made a bunch of mistakes:

  1. Made assumption about the cleanliness of my power supply. I am more used on digital design, I just sprinkled a bunch of capacitors, could not hear any noise on the sensor and called it a day. I've got a more professional supply on the way and will make sure to be more cautious with regards to decoupling capacitors
  2. My constant current circuit does not seems to be working properly. I need a voltage reference on the high side
  3. My quad op-amp should have its unused inputs tied.

I also made the formal mistake of not sufficiently relying on my scope and instead focusing on noise on the audio output. This stems from difficulty in characterising what is "problematic" noise from noise that is to be expected. If that is useful, here is a fft on the output;

  • When the photodiode is covered oscilloscope no led
  • When the photodiode is exposed oscilloscope led

EDIT: 11/12/23

In the end here what seemed to be the issue :

  • A driver circuit that was unstable with a single supply op amp. I am not totally sure why, but switching to dual supply solved some of the issue, even when I made sure not to run the op amp into the rails (it might also be that the dual supply op amp was less noise prone). That driving circuitry might also not be the best choice (also talked about here, and described as "prone to oscillate").
  • A bad power supply that added to the noise

The very same circuit with a dual supply was up to my expectations. I've still got to get a better grasp of op amps functionning (I'm working my way through the Art of Electronics, but slowly), but one thing that really limited me was my ability to characterize the noise with a scope. Audible difference did not show up significantly on the FFT, so it would really help if I could be directed to a primer on the matter. Anyway, RusselH's answer was the most complete based on my incomplete description so I'll mark it as accepted !

  • 2
    \$\begingroup\$ probably pickup of electromagnetic interference. Your amplifier might be more of a radio receiver than an optical receiver! You shouldn't be hearing a 50 Hz buzz even with an incandescent bulb (is your desk lamp incandescent?), but 100 Hz (the intensity goes from maximum to zero twice per sine period!). \$\endgroup\$ Nov 20, 2023 at 13:26
  • 2
    \$\begingroup\$ It's probably not 1/f if you can hear it with a gain-bandwidth product of just 1 MHz, by the way. \$\endgroup\$ Nov 20, 2023 at 13:27
  • 1
    \$\begingroup\$ if you can hear it on your headphones, if you don't have an oscilloscope, even the line-in of your PC/laptop would be a good start. \$\endgroup\$ Nov 20, 2023 at 13:27
  • 3
    \$\begingroup\$ Your transimpedance amplifier may be marginally stable and starting to ring when perturbed by a strong input. You could add a feedback capacitor and see if the problem goes away. Scope would give you a good idea if it's oscillating. \$\endgroup\$ Nov 20, 2023 at 13:49
  • 2
    \$\begingroup\$ If U1B pin5 is connected to "VCC" which is also UI1s supply then U1B is clipped, its output stuck to VCC, and the LED is simply driven by the power supply voltage. What are you using as power supply? Also (separate issue) you should AC couple the headphones if you don't want to risk burning the coil or the opamp. \$\endgroup\$
    – bobflux
    Nov 20, 2023 at 14:10

2 Answers 2

  1. The quad op-amp has two unused elements. Vout should be tied to IN-. IN+ should be biased to VCC/2 and capacitivily bypassed to VSS.
  2. I see no power supply decoupling. Quad and dual op-amps require significant capacitance between the VCC pin and the VSS pin to decouple the internal amplifiers from each other.
  3. Vin+ on both amplifiers should be kept between VCC-0.3V and VSS+0.3V. Otherwise power supply rejection is compromised.

The LED is powered through DC (I would say reasonably clean DC),

I would say, "Not so".

When I approach the LED, I hear a very strong background high frequency white noise (some sort of hiss).

This clearly indicates that there is a problem with the emitter or its current source.

The current source is not functioning as mentioned in the comments because the reference voltage for across R3 is VCC. The output of the op-amp cannot rise above VCC so the current through the LED will not be as expected and will fluctuate with the VCC. If the diode is barely turned on, it will be noisy.

Use a voltage reference (like the MAX6101) near 1.2V output to drive U1 pin 5. Adjust R3 for the required LED current. (Red to infrared are preferred).

so where is this "optical noise" coming from?

Hard to tell. The circuits need to be fixed first. I have given some possible sources, but no guarantees.

Is this the 1/f noise some papers talk about?

No. You would hear a popping sound. It is sometimes called flicker noise.

Is there a driver circuit to avoid it?

Proper circuit design and implementation first. A poor design cannot be fixed with a better component.

Alternative light sources?

Read the data sheets. Choose the one with the best noise characteristics.


So the 50Hz (and harmonics) sound IS there ONLY when the photodiode is NOT covered? Then it is responding to mains related variation in light intensity coming from your desklamp and other lighting. Other sounds are likely related to other light sources that contain AC component for one reason or another.

The membrane is likely modulating the light intensity, that's why its vibrations are picked up. (In fact you just invented a new form of microphone, it is now public domain ;-) )

Interesting that when you approach that has an effect. Reflected light from your body? Or some high frequency EM interference re-radiating via your body? I couldn't say but the body is mostly water and acts quite well as a sort of antenna.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.